CN1815979A - Method and apparatus for providing network connector - Google Patents

Abstract

The present application discloses a method, apparatus and program storage device for providing mutual failover and load balancing between interfaces in a network. An active virtual interface and a passive virtual interface are provided for each of the first and second computer interfaces. The active virtual interface of the first computer interface is communicatively coupled to the passive virtual interface of the second interface. The active virtual interface of the second computer interface is communicatively coupled to the passive virtual interface of the first computer interface. When the first and second computer interfaces are available, directing data flow through each of the active virtual interface of the first computer interface and the active virtual interface of the second computer interface, otherwise directing data flow Active virtual interface and passive virtual interface to available first or second interface.

Description

Apparatus and method for providing network connectivity

technical field

The present application relates generally to network computer systems and, more particularly, to methods, apparatus and program storage devices for providing mutual failover and load balancing between network interfaces.

Background technique

Computer systems linked to each other in a network are commonly used in businesses and other organizations. Networks of computer systems ("networks") provide many benefits to users, such as increased productivity, increased flexibility, and convenience, while also providing resource sharing and allocation.

Networks may be structured in different ways depending on details of the particular implementation, such as the hardware used and the physical location of the equipment, and depending on the particular purpose of the network. Generally, a network includes one or more server computer systems, each communicatively coupled to a number of client computer systems.

As the use of networked computer systems has increased, a need has arisen to provide additional bandwidth in order to handle electronic traffic on the network. For example, insufficient bandwidth may cause data congestion in the supply line between the client and the server. This blocking can significantly limit network performance.

A network interface card (NIC) is typically used to connect a server, or any computing device, to a network. Such NICs include, for example, Ethernet or Token Ring cards that plug into desktop computers or servers. The NIC implements physical layer signaling and media access control (Media Access Control, MAC) for computers attached to the network. Multiple NICs can effectively attach a computer to the network multiple times. Doing so proportionally increases the potential bandwidth to the network. Multiple NICs also provide resiliency and redundancy should one of the NICs fail. In the event of a failure of one NIC, one of the other NICs is used to handle traffic previously handled by the failed NIC, thereby increasing the reliability of the overall system. Therefore, there is a need to be able to detect that a NIC has failed, and when a failed NIC is detected, to be able to switch to a functioning NIC (this operation is referred to as fault tolerance and failover support). In a host operating system, a NIC is usually represented via a core object called a "network interface". Herein, a network interface directly used by Internet Protocol (IP) is referred to as an "IP interface". Also, the interface directly corresponding to the NIC is referred to as a physical interface. As described herein, an interface derived from a physical interface is called a logical or virtual interface.

Load balancing is a technique used to reduce data bottlenecks caused by overloaded communication networks. In load balancing, traffic between servers and the network is shared over multiple NICs. This load balancing usually requires specialized software. Load balancing also provides fault tolerance, which is used to maintain data communication between the server and the network in the event of an interruption in the data link. When a link fails, the load is failed over to a backup or auxiliary link so that signal continuity is maintained.

A well-known technique is to group multiple physical links together, thereby making them appear to be a single network interface to the Internet Protocol (IP) layer of the TCP/IP stack. Load balancing and failover are then implemented among the links without the knowledge of the IP layer. Examples of such techniques are 'bonding' drivers in Linux, Etherchannel or the IEEE802.3ad link merging standard.

However, these techniques suffer from several drawbacks. For example, load balancing is implemented below the IP layer because the system treats multiple physical links as a single NIC. In other words, a large number of NICs exist as a single interface to the IP protocol. Therefore, network layer information such as routing tables cannot be used for load balancing of data traffic. Common tools that work at the network layer are also not applicable. These disadvantages also exist in failover mode.

Link merging techniques as described above also require specialized switches capable of treating multiple switch ports as one; in the case of directly connected peer-to-peer systems, both endpoints must be configured to support the standard . Furthermore, failure of the switch takes connectivity out of all links. In an alternate mode that only supports failover but not load balancing, links can be connected to different switches, however in this configuration only one link can be active at a given time.

Load balancing can also be provided at the IP layer, where the load is balanced among multiple IP interfaces. The data load is balanced according to routing table entries pointing to the particular IP interface for a given route. When a NIC fails, an alternate method for failover must be implemented, and time must be spent updating the routing tables. The link-level failover described earlier occurs in very short (milliseconds) intervals, whereas route propagation can take much longer. Additionally, an IP address needs to be associated with the backup interface and the failover MAC address notified to the peer. Therefore, there is a need for a method of achieving load balancing at the IP layer while providing fast failover.

It can be seen, then, that there is a need for methods, apparatus and program storage devices for providing mutual failover and load balancing between interfaces at the IP layer of a network.

Contents of the invention

To overcome the limitations described above, and to overcome other limitations which will be more apparent upon reading and understanding this specification, the present invention discloses a method for providing mutual failover and Method, apparatus and program storage device for load balancing.

Embodiments of the present invention solve the above problems by combining active and passive virtual interfaces with cooperating computer interfaces to share bandwidth when all computer interfaces are active, and in the event that one of the computer interfaces failure, then provide failover. The present invention also provides a load balancing strategy to be determined at the IP layer without requiring or using specialized hardware or counterparts capable of implementing the same features in peer-to-peer systems. The IP layer can implement load balancing among multiple interfaces while avoiding network card failures.

A system in accordance with the principles of the present invention includes at least first and second computer interfaces having a common set of one or more identifiers for accommodating a plurality of MAC addresses; first active and passive virtual interfaces capable of communicating ground coupled to the first computer interface, and a second active and passive virtual interface communicatively coupled to the second computer interface, wherein the first active virtual interface is virtual to the second passive virtual interface combined, and the second active virtual interface is virtually combined with the first passive virtual interface; and at least one access controller, coupled to the virtual interface, for controlling flow via the communicatively coupled virtual interface to data flow of at least first and second computer interfaces, directing data flow via said first and second active virtual interfaces when said first and second computer interfaces are available, and otherwise, when said first Or when the second computer interface is unavailable, direct the data flow to the active virtual interface and the passive virtual interface of the available first or second interface.

In another embodiment of the present invention, a program storage device is provided, which can be read by a computer and can contain one or more program instructions executable by a computer, so as to execute the Operation of data streams interfaced with a second computer, thereby providing mutual failover and load sharing. The operations include providing an active virtual interface and a passive virtual interface to each of the first and second computer interfaces; communicatively coupling; communicatively coupling the active virtual interface of the second computer interface with the passive virtual interface of the first computer interface; and when the first and second computer interfaces are available, via the the active virtual interface of the first computer interface and the active virtual interface of the second computer interface to direct data flow, otherwise when one of the first or second computer interface is not available, the data flow is directed to Active virtual interface and passive virtual interface of the first or second interface available.

In another embodiment of the present invention, a method for supporting data flow at a first and second computer interface to provide mutual failover and load sharing is provided. The method includes providing an active virtual interface and a passive virtual interface for each of the first and second computer interfaces; connecting the active virtual interface of the first computer interface to the passive virtual interface of the second interface communicatively coupling; communicatively coupling the active virtual interface of the second computer interface with the passive virtual interface of the first computer interface; and when the first and second computer interfaces are available, via the the active virtual interface of the first computer interface and the active virtual interface of the second computer interface to direct data flow, otherwise when one of the first or second computer interface is not available, the data flow is directed to Active virtual interface and passive virtual interface of the first or second interface available.

In another embodiment of the present invention, a system for providing network connections for mutual failover and load sharing is provided. The system includes means for providing an active virtual interface and a passive virtual interface for each of the first and second computer interfaces; for connecting the active virtual interface of the first computer interface to the second interface means for communicatively coupling the passive virtual interface of the second computer interface; means for communicatively coupling the active virtual interface of the second computer interface with the passive virtual interface of the first computer interface; and means for when the directing data flow via each of the active virtual interface of the first computer interface and the active virtual interface of the second computer interface when the first and second computer interfaces are available, otherwise directing the data flow to the available Active virtual interface and passive virtual interface device of the first or second interface.

These and various other advantages and features which characterize the invention are pointed out with particularity in the claims annexed to and forming a part of this application. However, for a better understanding of the invention, its advantages obtained by its use and objects, reference should be made to the accompanying drawings forming a part hereof, and to the accompanying description, in which there are illustrated and described specific examples of apparatus in accordance with the invention.

Description of drawings

Referring now to the drawings, wherein like reference numerals indicate corresponding parts throughout:

Figure 1 illustrates a system network hierarchy describing the state of the art;

Figure 2a is a diagram of a network with two interfaces A and B to provide failover and load balancing in accordance with an embodiment of the present invention;

Figure 2b shows the network of Figure 2a in failover mode;

Fig. 3 illustrates using a virtual IP address to manage network interface access control according to an embodiment of the present invention;

Figure 4 illustrates a method for providing network connectivity for mutual failover and load sharing according to an embodiment of the present invention; and

FIG. 5 illustrates a system for performing operations to support data flow at a first and second computer interface to provide mutual failover and load sharing, in accordance with an embodiment of the present invention.

Detailed ways

In the ensuing description of the embodiments, reference is made to the accompanying drawings which form a part hereof, and in which is shown by way of illustration specific embodiments in which the invention may be practiced. It is to be understood that other embodiments may be utilized as structural changes may be made without departing from the scope of the present invention.

The present invention provides a method, apparatus and program storage device for providing mutual failover and load balancing between interfaces in a network. The present invention combines active and passive virtual interfaces with cooperating computer interfaces to share bandwidth when all computer interfaces are working, and provides failover in the event that one of the computer interfaces fails.

A single physical interface can be used to create a virtual interface. Two or more virtual interfaces can be combined to form a single interface that distributes or load balances its packets among the combined virtual interfaces. In the context of the present invention, the term 'join' or 'combination' is used to logically combine two virtual interfaces so that they can be used as a single IP interface. Here, such a combined virtual interface formed by combining one or more virtual interfaces is called a component interface. With appropriate controls, a component interface can send and receive its packets via the physical interface from which it originated.

Figure 1 illustrates a system network hierarchy 100 describing the prior art. Failover and load balancing mechanisms can be provided by configuring the device to be based on a combination of two physical interfaces. The hierarchy of FIG. 1 shows an IP interface C 100, which is formed by two NICs a 140 and b 130. In the operating system's network stack, these two NICs are represented as physical interface A 110. All data sent along IP interface C 100 can be load balanced in two interfaces B 120 and A 110. The media independent interface (MII) signal failure is used to detect that a NIC is not working. This detection and failover is instantaneous (measured in milliseconds).

It should be understood that a composition may include multiple component interfaces and is not limited to the two component interfaces B 120 and A 110 as shown in FIG. 1 . Additionally, combinations of component interfaces such as component interface B 120 and A 110 may be configured such that only one of the component interfaces, such as component interface A 110, is active at all times, while the other component interface, such as component Interface B 120 acts as a passive backup. When a failure is detected on the first component interface A 110, a backup interface such as component interface B 120 may become active. According to this configuration, each of interfaces b 130 and a 140 is connected to a separate switch. However, this configuration wastes the resources of the passive backup since it remains idle.

The present invention proposes a method for providing mutual backup among interfaces thereby preventing any resource from being idle. As described above, fast failover (eg, in milliseconds) can be provided by using MII detection at the device. As will be seen, embodiments of the present invention continue to balance this fast failover based on MII monitoring, because when an interface failure is detected, the passive component interface on the functioning physical interface is immediately made active.

In accordance with link pooling implementations such as trunking, channel bonding, or IEEE's 802.3ad standard, multiple links appear to be a single IP interface while providing load balancing and failover of data across the links. In a link consolidation configuration, multiple network links appear as a single link with a single logical network interface (which appears to be an IP interface) and has one virtual Media Access Control (MAC) address. The MAC address of one of the interfaces belonging to the merged link provides the virtual address of the logical link. In a typical link consolidation configuration, the data can be balanced across the physical links based on basic algorithms such as round robin routing of source and destination MAC addresses, logical XOR operations, and the like. Since multiple links on different switch ports are considered to be part of the same logical link, specialized switches are required that are able to treat a group of ports as a single port and load balance traffic among these ports.

Figure 2a is a diagram of a network 200 having two interfaces A 215 and B 230 to provide failover and load balancing in accordance with an embodiment of the present invention. Both interfaces A 215 and B 230 are divided into two parts, for example, A 215 is divided into Ax 205 and Ay 210, while B 230 is divided into Bx 220 and By 225. Interfaces A 215 and B 230 correspond to component interfaces A 110 and B 120 in FIG. 1 . IP interfaces E1 201 and F1 202 are formed. E1 201 is formed by combining virtual interfaces Ax 205 and Bx 220. F1 202 is formed by combining virtual interfaces Ay 210 and By 225. Each combination of E1 201 and F1202 supports active and passive component interfaces. Thus, for interface E1 201, Ax 205 is an active component and Bx 220 is a passive component. For the interface F1 202, By 225 is the active component and Ay 210 is the passive component.

If one of the physical interfaces A 215 or B 230 becomes unavailable, the component interface may redirect data flow to the operational physical interface via the component interface associated with the operational physical interface. Thus, the two interfaces A 215 and B 230 can remain active, act as backups for each other and/or share data packets in both interfaces. Load balancing can operate under normal operating conditions when both interfaces are active. Data sent on combo interface E1 201 will flow through Ax 205 on physical interface A 215.

Likewise, under normal conditions, data sent on combo interface F 1202 will flow through By 225 on physical interface B 230. Each interface is connected to a common OS1 layer 2 destination, such as switch 240 . Use combined interfaces E1 201 and F1 202 as IP interface. With the help of network layer techniques, such as equal path routing, policy based routing or other load balancing techniques, the network load can be balanced among interfaces E1 201 and F1 202 while taking full advantage of the resources provided by both interfaces A 215 and B 230 physical bandwidth. As mentioned above, the data actually flows through the active components of the combination and then finally through the physical interface. The aggregate throughput across the interface is thus maximized, while providing mutual backup for interface failures, while providing network-level control of load balancing.

Figure 2b illustrates the network 200 of Figure 2a when one of the physical interfaces B 230 fails. When either interface A 215 or interface B 230 is unavailable, failover is initiated and the available interface will handle the load for both A 215 and B 230 interfaces. As shown in Figure 2a, the active component Ax 205 of the active interface A 215 is used to normally receive data packets via E1 201. However, the passive component Ay 210 of the active interface A 215 receives a data packet originally intended for the unavailable interface B 230 via the combined interface F1 202, which redirects the data packet to Ay 210 instead of By 225. This failover is not detected by the IP layer, so the IP layer continues to work as before. Thus, fast failover can be performed without disrupting IP load balance. Since the IP interfaces E1 201 and F1 202 remain unaffected when physical interface A 215 or B 230 fails, there is no routing update requirement. The IP layer can thus be used to load balance between the two interfaces using equal paths, policy based routing, or using various network level tools.

Communication with routers/peers that received packets through the failed NIC must continue. Three methods for maintaining communication when, for example, a NIC fails, will be described here. A network interface card (NIC) for simultaneously supporting multiple MAC addresses is used. The same set of MAC addresses is assigned to all physical interfaces that end up forming a combination, such as interfaces A 215 and B 230. Each MAC address is associated with a respective IP address for use at the IP interface E1 201 or F1 202. The MAC address of the outgoing packet is selected based on the source IP address. As a default, only one component is active on a given physical interface, component Ax 205, while Ay 210 is passive on interface A 215. Data is normally transferred from the IP interface through the component interface. Accordingly, each physical interface, such as interfaces A 215 and B 230, is assigned the MAC address corresponding to the active component as its primary MAC, and advertises this to switch 240 through an appropriate mechanism. When enabled, the physical interface is configured to filter out all packets with a destination MAC address not equal to the primary MAC address of the physical interface. Normal communication between IP nodes ensures that the switch tables are updated with the MAC addresses associated with their ports.

The method of passing the MAC address to the switch 240 connected to interfaces A 215 and B 230 when failover is to transmit the source MAC set to the failed NIC's primary MAC and the destination MAC set to the active NIC interface's primary MAC Ethernet frame. The payload of the transmitted frame conforms to the Ethernet protocol requirements, but the data is all zeros. In this embodiment, the switch 240 will update its table to indicate the location of the MAC on the receiving port. The frame's "Ethertype" value SHOULD be set to a value not supported on the system, or to a special value requested by the Internet Assigned Number Authority (IANA), so that if the packet is sent back to NIC, it is filtered out by the driver. In this embodiment of the invention, upon failover, the MAC address associated with the IP address is moved to the functioning NIC without having to notify all IP clients/peers of the failover. The functioning NIC is suitably configured to also receive packets on the failover MAC address.

The MAC address of the interface to which packets on a local area network (LAN) are transmitted. IP implements address resolution techniques in order to determine the MAC address corresponding to a peer's IP address. An address resolution response to any particular IP address, via ARP or Neighborhood Discovery, always returns the primary MAC of the physical interface associated with the active component of the IP interface bound to it, e.g., IP interface E1 201 returns A 215 primary MAC address. When an interface fails, the passive components of the active physical interface A 215 or B 230 become active through software control. Additionally, Ethernet frames as described above are transmitted on the physical interface of the primary MAC with the failed physical interface.

While multiple MACs per NIC are not supported, another method can be used to maintain communication when a NIC fails. In this method, a gratuitous ARP can be transmitted to announce that the IP address is now associated with the MAC address of the functioning NIC.

Furthermore, when the NIC does not support multiple MAC entries, when the NIC fails, the third method can be used to maintain communication. In receiving IP nodes of the same network, independent multicast MAC values can be assigned to physical NICs as their MAC addresses, because of the check of whether the source MAC address in the received Ethernet frame is a multicast address. When failing over, the multicast MAC address of the failed NIC can be added to the multicast list on the active interface. Thus, all packets received on that multicast address will be received by the NIC.

FIG. 3 illustrates the use of virtual IP addresses 360 to manage network interface access control in accordance with an embodiment 300 of the present invention. A virtual IP address is an address that is not associated with a physical IP interface, but a logical address that can be logically mapped to a physical IP address. Since this address is not affected by the state of the physical interface such as the interface being recorded, it will always exist. Such an IP address can be configured to be on a network that appears to be reachable only through virtual bonded interfaces B1 301 and B2 302. With this setup, all communications to/from the system can use the virtual IP address; however the data itself can be load balanced across the IP routes associated with B1 301 and B2 302. Such control occurs at the network layer, layer 3, and provides failover and load balancing without requiring specialized hardware, thereby providing better control than link consolidation schemes. IP address 360 provides IP routing (e.g. equal path) via two IP addresses IPa 350 and IPb 355 via virtual interfaces B1 301 and B2 302 and virtual interfaces Ax 305, Ay 310, Bx 320 and By 325 are coupled to interfaces A 315 and B 330 to provide load balancing between the two physical interfaces A 315 and B 330. Interfaces A 315 and B 330 are communicatively coupled to a network device 340, which may include a switch or other IP node. MII monitoring is used to detect failures when the interface becomes passive, and passive components in the combined IP interface become active. The failure detection generates gratuitous ARP or Ethernet frames to inform the switch that packets from the same IP address flow through different ports. However, routing table entries, including routing table entries at other routers in the network, do not require any modification because no IP interfaces are affected. For example, if interface A 315 fails, then passive virtual interface Bx 320 becomes active and uses a gratuitous ARP or Ethernet frame to inform switch 340 that packets from the same IP address, IPa 350, are now communicating with the interface The port to which the B 330 is connected is received.

It should be understood by those skilled in the art that the present invention can be extended to multiple NICs for providing load balancing and mutual backup for mutual failover. For example, each NIC can contribute to N component interfaces. Each IP interface is formed using one virtual component interface from each NIC that has only one of the active components. Other component interfaces are passive. When a NIC fails, the IP interface with active components from that NIC will fail over to one of the other component interfaces associated with the NIC that is still active. Embodiments of the present invention can also be extended to include a pool of backup component interfaces that are only combined with the combined interface if the active component interface of the IP interface fails.

FIG. 4 illustrates a method 400 for providing network connections for mutual failover and load sharing according to an embodiment of the present invention. In accordance with the method, at 410, an active virtual interface and a passive virtual interface are provided for each of the first and second computer interfaces. At 420, the active virtual interface of the first computer interface and the passive virtual interface of the second interface are communicatively coupled. At 430, the active virtual interface of the second computer interface and the passive virtual interface of the first computer interface are communicatively coupled. At 440, a determination is made as to whether the first and second computer interfaces are available. At 445, when it is determined to be available, then at 450, data flow is directed via each of the active virtual interface of the first computer interface and the active virtual interface of the second computer interface. Otherwise, at 455, when one of the first and second computer interfaces is not available, then at 460, data flow is directed to the active virtual interface and the passive virtual interface of the available first or second interface.

FIG. 5 illustrates a system 500 according to the present invention, wherein the processes illustrated in accordance with FIGS. 1-3 may be embodied on a computer-readable medium or carrier, such as one or more of the Fixed and/or removable data storage device 568, or other data storage or data communication device. Computer programs 590 representing processes embodied on removable data storage device 568 may be loaded into memory 592 or into controller system 500, such as processor 510, to configure controller system 500 of FIG. 5 for execution. The computer program 590 includes instructions which, when read and executed by the controller 500 of FIG. 5 , cause the controller system 500 to perform the steps necessary to implement the steps or elements of the present invention.

The foregoing description of exemplary embodiments of the present invention has been presented for purposes of illustration and description. It is not intended to be exhaustive or to limit the invention to the precise forms disclosed. Many modifications and variations are possible in light of the above teachings. It is intended that the scope of the invention be limited not by this detailed description, but by the appended claims.

Claims (25)

1. An apparatus for providing network connectivity for mutual failover and load sharing, comprising: at least first and second computer interfaces having a common set of one or more identifiers for accommodating a plurality of MAC addresses; A first active and passive virtual interface communicatively coupled to the first computer interface, and a second active and passive virtual interface communicatively coupled to the second computer interface, wherein the first an active virtual interface is virtually bonded to the second passive virtual interface, and the second active virtual interface is virtually bonded to the first passive virtual interface; and at least one access controller coupled to the virtual interface for controlling data flow to the at least first and second computer interfaces via the communicatively coupled virtual interface, when the first and second directing data flow through said first and second active virtual interfaces when a computer interface is available, otherwise directing data flow to an available first or second computer interface when one of said first or second computer interfaces is not available An active virtual interface and a passive virtual interface for one or the second interface. 2. The apparatus of claim 1, further comprising a switch coupled to the at least first and second computer interfaces, the switch configured to process Ethernet frames for the same source and destination. 3. The device of claim 1, wherein the first and second computer interfaces comprise at least two network interface cards. 4. The apparatus of claim 3, wherein each network interface card supports multiple MAC addresses simultaneously. 5. The device of claim 1, wherein the computer interface is configured to transmit a frame to provide failover. 6. The apparatus of claim 5, wherein the frame transfers a source set to the primary MAC of the failed NIC and a destination MAC set to the primary MAC of the active NIC interface. 7. The device of claim 1, wherein the first active virtual interface and the second passive virtual interface form a first component interface, and the second active virtual interface and the first passive The source virtual interface forms the second component interface. 8. The apparatus of claim 1, wherein the access controller determines the availability of at least the first and second computer interfaces by sending a gratuitous Address Resolution Protocol (ARP). 9. The device of claim 1, wherein at least a portion of the virtual interfaces are selected from a pool of virtual interfaces. 10. The device of claim 1, wherein the access controller assigns a virtual internet protocol address to each of at least the first and second interfaces and routes the data flow using internet protocol routing. 11. The device of claim 1 , wherein independent multicast MAC values are assigned to the first and second computer physical interfaces as corresponding MAC addresses, and, upon failover, the failed computer interface is is added to the multicast list on the remaining computer interfaces to ensure that all received packets with the multicast MAC address of the failed computer interface are received by the remaining computer interfaces. 12. A method for providing network connectivity for mutual failover and load sharing, comprising: providing an active virtual interface and a passive virtual interface for each of the first and second computer interfaces; communicatively coupling the active virtual interface of the first computer interface with the passive virtual interface of the second interface; communicatively coupling the active virtual interface of the second computer interface with the passive virtual interface of the first computer interface; and directing data flow through each of the active virtual interface of the first computer interface and the active virtual interface of the second computer interface when the first and second computer interfaces are available, otherwise when the first computer interface When one of the first or second computer interfaces is unavailable, data flow is directed to the active virtual interface and the passive virtual interface of the first or second interface that is available. 13. The method of claim 12, further comprising sending a gratuitous ARP to determine the availability of one of the first and second interfaces. 14. The method of claim 12, wherein directing data flow includes directing data flow using active virtual interfaces and passive virtual interfaces of the first and second computer interfaces. 15. The method of claim 14, wherein at least a portion of the virtual interfaces are selected from a pool of virtual interfaces. 16. The method of claim 12, wherein the step of providing each of the first and second computer interfaces with an active virtual interface and a passive virtual interface further comprises providing at least one of the first and second computer interfaces with Two network interface cards. 17. The method of claim 16, wherein the step of providing at least two network interface cards to the first and second computer interfaces further comprises providing support for multiple MAC addresses simultaneously. 18. The method of claim 12, further comprising configuring the computer interface to transmit a frame to provide failover. 19. The method of claim 18 , wherein the step of configuring the computer interface to transmit a frame to provide failover further comprises: using the frame to convey a source set to the primary MAC of the failed NIC and set to Destination MAC of the primary MAC of the active NIC interface. 20. The method of claim 12, wherein the step of directing data flow comprises: assigning a virtual internet protocol address to each of the at least first and second interfaces; and The data flow is routed using Internet Protocol routing. 21. The method of claim 12, wherein the step of providing each of the first and second computer interfaces with an active virtual interface and a passive virtual interface further comprises: using the first active virtual interface and the second computer interface. Two passive virtual interfaces are used to form a first component interface, and the second active virtual interface and the first passive virtual interface are used to form a second component interface. 22. The method of claim 12, wherein the step of directing data flow further comprises determining the availability of at least the first and second computer interfaces by sending a gratuitous Address Resolution Protocol (ARP). 23. The method of claim 12, wherein the step of directing data flow further comprises: assigning independent multicast MAC values as corresponding MAC addresses to the first and second computer physical interfaces, and upon failover , adding the multicast MAC address of the failed computer interface to the multicast list on the remaining computer interfaces to ensure that all received packets with the multicast MAC address of the failed computer interface are received by the remaining computer interfaces. 24. A system for providing network connectivity for mutual failover and load sharing, comprising: means for providing an active virtual interface and a passive virtual interface for each of the first and second computer interfaces; means for communicatively coupling the active virtual interface of the first computer interface with the passive virtual interface of the second interface; means for communicatively coupling the active virtual interface of the second computer interface with the passive virtual interface of the first computer interface; and for directing data flow through each of the active virtual interface of the first computer interface and the active virtual interface of the second computer interface when the first and second computer interfaces are available, and otherwise Means that flow is directed to the active virtual interface and the passive virtual interface of the available first or second interface. 25. A program storage device comprising instructions for performing the method of any one of the preceding method claims.

Images